Choose the asset before comparing service proposals. Record its type, age, insulation, operating history, and voltage class. A 15 kV cable, a 132 kV transformer, and medium-voltage switchgear need different test setups and interpretation. IEEE Std 400.3-2022 addresses field partial-discharge testing of shielded power cable systems rated 5 kV and above. That threshold matters. Ask whether the provider’s method suits the actual asset, not just its nominal voltage. Note accessories, joints, and terminations; defects can hide there.

Define the objective in plain terms: establish a baseline, investigate a suspected defect, or track change over time. IEC 60270:2000+A1:2015 describes apparent-charge measurement in picocoulombs, but a single pC value does not explain defect severity by itself. Request test voltage, sensor type, noise controls, calibration approach, and site conditions alongside results. For example, a noisy substation may make weak signals harder to separate from interference. Ask for raw patterns and a clear account of uncertainty. A polished report can still answer the wrong question. It is easy to over-specify the test before agreeing what decision the result must support.

Choosing partial discharge analysis means checking how the service measures and reports apparent charge. IEC 60270:2000+A1:2015 defines apparent-charge measurement for electrical discharge testing; the unit is picocoulombs (pC), where 1 pC equals 10⁻¹² coulombs. Ask whether the test team calibrates the measuring circuit before readings and records test voltage, sensor position, and background noise. A number without these details can mislead. IEEE 400.3-2022 also offers guidance for partial-discharge testing of shielded power cables.

Compare results under consistent conditions, not by pC alone. For example, a 25 pC reading is illustrative, not a universal pass-or-fail threshold. Limits depend on equipment type, test setup, and operating context; IEC 60270 does not provide one acceptance limit for every asset. Request the raw readings, calibration details, and a clear explanation of uncertainty. A skilled analyst should distinguish repeatable phase-related activity from electrical interference, then explain what the finding means for that specific asset. Field conditions are rarely perfect; even experienced teams may need to repeat a measurement.

Choosing partial discharge analysis services means checking how well the provider matches its method to your equipment. IEC 62478 covers electromagnetic and acoustic approaches, including UHF measurement. UHF sensors detect fast electromagnetic signals, often inside gas-insulated switchgear. The metal enclosure can help contain signals, but sensor placement and enclosure design affect what reaches the instrument. A quiet-looking trace is not proof of healthy insulation. Noise matters.

Acoustic sensors pick up pressure waves from discharge activity. A technician may compare arrival times across several sensors to estimate a fault’s location. This can be useful when access points are limited, though sound weakens as it travels through barriers and joints. Pumps, fans, and loose panels can also complicate readings. Each method has blind spots. Ask whether the service team records sensor positions, operating conditions, background noise, and repeat measurements. Those details make results easier to review later. A careful provider should explain uncertainty, not just present a colored graph. Ideally, UHF and acoustic findings are cross-checked when practical; they do not always agree, and that disagreement deserves investigation. During a site visit, ask how the team distinguishes recurring PD patterns from interference. A polished report helps, but clear evidence and honest caveats matter more.

Choosing partial discharge analysis services means checking how results are produced, not just how sensitive the instrument is. Ask whether calibration uses traceable references and covers the sensor, cable, and acquisition chain used onsite. A certificate alone is not enough; confirm its date, uncertainty, and frequency range. Small differences matter.

At 50/60 Hz, ask how the service distinguishes phase-related discharge pulses from switching transients, radio interference, and nearby equipment noise. Request a noise-rejection demonstration under realistic site conditions, including the planned grounding and sensor placement. A clean laboratory trace can look convincing, yet industrial spaces rarely stay quiet. Test it onsite. Check whether reports record rejected events and explain the criteria, rather than simply hiding interference. No test is perfect; field noise can still mask weak events.

Sensitivity should be stated as a measurable detection threshold, with sensor type, bandwidth, test setup, and noise floor disclosed. Lower thresholds help only when readings remain repeatable and distinguishable from background noise. Ask for checks at several signal levels, especially near the stated limit, and compare repeated captures. If thresholds shift with cable length or frequency, the provider should explain and document the limitation. One weakness to watch for is a precise specification with vague operating conditions. That gap deserves a question.